21st-ai
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill interacts with the user's environment by calling the 21st CLI for UI generation and iteration tasks.
- Evidence: The skill utilizes commands such as
21st generate,21st iterate, and21st taketo manage UI drafts. - [INDIRECT_PROMPT_INJECTION]: The skill processes UI code and specifications generated by an external AI service, presenting a standard ingestion surface for third-party content.
- Ingestion points: Data enters the agent's context through the output of the
21st takecommand (HTML or copy-prompts). - Boundary markers: The instructions explicitly advise treating the output as a "design spec" and rebuilding it rather than pasting raw code verbatim.
- Capability inventory: The skill is intended to help the agent "pull the final code into the project," which involves file-writing capabilities.
- Sanitization: The skill implements mitigation by guiding the agent to interpret the results as design instructions rather than executable code.
- [EXTERNAL_DOWNLOADS]: The skill is built around the use of the
@21st-dev/clipackage, which is the official tool from the skill author.
Audit Metadata